Multiplex communications – Pathfinding or routing – Combined circuit switching and packet switching
Reexamination Certificate
1997-12-12
2002-03-12
Yao, Kwang B. (Department: 2664)
Multiplex communications
Pathfinding or routing
Combined circuit switching and packet switching
C370S468000
Reexamination Certificate
active
06356545
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Technical Field
The present invention relates generally to an Internet telephone system operating over a Public Switched telephone Network (PSTN), and more specifically to an Internet telephone system with codecs that dynamically change from packet to packet.
2. Discussion of the Prior Art
The idea of sending telephone calls over the Internet is relatively new, brought on by the desire to avoid expensive long distance telephone bills from the major telephone companies. While the concept of sending audio and video data, e.g. cable programming, over the Internet has been known since 1990, it was not until 1993 that a program called Maven was created to transmit voice data using a personal computer. In 1995, VocalTec offered a beta test version of its Windows-based Internet telephone, and that same year DigiPhone offered full duplex Internet telephone system, which allowed users to talk and listen simultaneously.
Several problems must be addressed to make an Internet telephone product commercially successful. One of the most important is maintaining sound quality despite dropouts or gaps caused by the Internet. The digital nature of the Internet has theoretical advantages vis-a-vis analog networks, but when the Internet is busy a caller may have difficulty getting through to another party. Moreover, since the Internet is built to transfer data packets rather than continuous streams of sound, there may be delays and losses.
For a telephone call to be placed over the Internet, the analog voice information must be converted into a digital format as a series of data packets that are communicated through the Internet's web of computers, routers and servers. Data compression algorithms are designed to prevent the customer from noticing delays between packets in the data stream.
Analog voice messages spoken by customers are digitized and then compressed by a compression/decompression (‘codec’) algorithm. There are at least ten different types of codecs, each designed to compress data optimally for a particular application. Some codecs use audio interpolation to fill in dropouts or gaps. Other codecs create high quality sound, but use complex algorithms that are slower to execute on a given computer. Still other codecs use faster compression algorithms, but the sound quality is not as high. Whether the speed of the compression algorithm is important for a particular application may depend on the speed of the computer executing the algorithm. Different codecs use different compression ratios to compress data. For example, one codec might compress data by a factor of two from 8 kbits/second to 4 kbits/second, while another codec might compress data by a factor of five from 8 kbits/second to 1.65 kbits/second. Codecs exist that have data compression factors of twelve, and even as high as fifty, but these require more complex mathematical algorithms and the resultant sound quality may depend on such things as the frequency and computer connection. Exemplary codecs include GSM, a European standard having a 5:1 compression ratio, and the TrueSpeech codec (of DSP Group, Santa Clara, Calif.) having a 15:1 compression ratio.
Prior art systems typically run only one codec at a time, although the codec may be specified initially by the user through adjustment of the computer settings or through selecting the codec from a file menu. Codec programs at both ends of an exchange must be able to understand each another, so compatibility between codecs may also be an issue.
Codecs do not address data dropouts or loss. Ordinarily, the Internet Protocol uses an Automatic Repeat Request (ARQ) to request retransmissions of lost messages, but voice transmission systems attempt to interpolate lost data rather than resend it.
It is generally known in prior art Internet telephone systems that codecs can be manually selected since both parties must be using the same codec to understand one another. U.S. Pat. No. 5,539,908 to Chen et al discloses a system for dynamically linking codec algorithms between file formats. While Chen et al supports a plurality of existing and future codec installations, codecs are only changed between file formats and not on a packet-to-packet basis.
U.S. Pat. No. 5,394,473 to Davidson discloses a device for coding and decoding of audio signals that optimizes between time and frequency resolution through selection of the coder. However, Davidson is not concerned with an Internet telephone system, nor changing the codec on a packet-to-packet basis.
The television and radio industries employ speech compression techniques in advertisement spots to minimize the amount of advertising time paid for by sponsors. Such techniques speed up the audio data while maintaining a constant pitch or frequency for the voiceover. U.S. Pat. No. 5,583,652 to Ware provides a technique known as time domain harmonic scaling for variable speed playback of an audio/video presentation while keeping the audio and video synchronized as well as the audio pitch undistorted. U.S. Pat. No. 5,555,447 to Kotzin et al mitigates speech loss in a communication system by buffering time-compressed speech in a FIFO until the FIFO is substantially empty. Thereupon, Kotzin et al transitions the communication system from time-compressed to normal speech.
SUMMARY OF THE INVENTION
The present invention sets forth a novel Internet telephone system architecture for providing full duplex operation with low voice latency. The architecture enables a dynamic change of codec from packet to packet in the same voice data stream in order to adapt to changing network conditions. The change of codec operates concurrently with a change in other factors including the level of redundancy of the error correction, the packet size and packet bundling. The architecture thereby seeks to attain the best speech quality and lowest latency given the level of data loss over the Internet detected by the system.
One further feature of the present invention is a technique for eliminating dead air space in the data stream by speeding up or slowing down the data from the buffer while maintaining a constant voice pitch.
REFERENCES:
patent: 4864562 (1989-09-01), Murakami et al.
patent: 5187591 (1993-02-01), Guy et al.
patent: 5394473 (1995-02-01), Davidson
patent: 5533004 (1996-07-01), Jasper et al.
patent: 5539908 (1996-07-01), Chen et al.
patent: 5555447 (1996-09-01), Kotzin et al.
patent: 5583652 (1996-12-01), Ware
patent: 5617423 (1997-04-01), Li et al.
patent: 5699369 (1997-12-01), Guha
patent: 5742773 (1998-04-01), Blomfield-Brown et al.
patent: 5881234 (1999-03-01), Schwob
patent: 5890108 (1999-03-01), Yeldener
patent: 5933803 (1999-08-01), Ojala
patent: 5940479 (1999-08-01), Guy et al.
patent: 6026082 (2000-02-01), Astrin
patent: 6052391 (2000-04-01), Deutsch et al.
patent: 6064653 (2000-05-01), Farris
patent: 6130883 (2000-10-01), Spear et al.
Chang Jerry
Vargo Mike
Carr & Ferrell LLP
Clarent Corporation
Yao Kwang B.
LandOfFree
Internet telephone system with dynamically varying codec does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Internet telephone system with dynamically varying codec, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Internet telephone system with dynamically varying codec will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2875750